The effects of different parameters on the number of haploid genomes screened.

<p>(A–C) Plots assuming all <i>F</i>1 animals are placed together on a single plate. <i>p</i> = 0.75, unless otherwise indicated. All plots were generated using the program Mathematica 5.0 (Wolfram Research). (A) Black line, graph of <i>n_act</i>/<i>N vs.</i> log₁₀<i>y</i> for <i>N</i> = 1,000. Blue and red lines, graphs of asymptotes and their equations. (B) Graphs of <i>n_act</i>/<i>N vs.</i> log₁₀<i>y</i>. Different colors indicate <i>n_act</i>/<i>N</i> for different specified values of <i>N</i>. Black line, Poisson approximation, colored lines, exact solutions. Inset, magnification of the graph for the region of log₁₀<i>y</i> between 0.3 and 0.5. For each value of <i>N</i>, <i>y</i> can only take on values such that 1/(<i>N</i>−1)≤<i>y</i>≤<i>N</i>−1. Furthermore, although for illustration purposes we have drawn the curves as continuous, <i>y</i> is <i>not</i> a continuous variable, and treating it as such only works for large <i>N</i>. This is most obvious for <i>N</i> = 10 where <i>y</i> can only take on the values 1/9, 1/4, 3/7, 2/3, 1, 3/2, 7/3, 4, and 9. (C) Graphs of <i>n_act</i>/<i>N vs.</i> log₁₀<i>y</i> for varying values of <i>p</i>, as defined in the text. Graphed using the Poisson approximation. (D) Graphs depicting the fractional error incurred when using the Poisson approximation to estimate <i>n_act</i>/<i>N</i> for screens in which one <i>F</i>1 animal is plated per plate. Although graphs are continuous, only integer values of <i>y</i> are relevant. Also note that the smallest allowable value of <i>y</i> is chosen so that at least 1 <i>F</i>2 animal is chosen per plate.